Deep Learning-Augmented Head and Neck Organs at Risk Segmentation From CT Volumes
Purpose: A novel deep learning model, Siamese Ensemble Boundary Network (SEB-Net) was developed to improve the accuracy of automatic organs-at-risk (OARs) segmentation in CT images for head and neck (HaN) as well as small organs, which was verified for use in radiation oncology practice and is there...
Guardado en:
| Autores principales: | , , , , , , , , , , |
|---|---|
| Formato: | article |
| Lenguaje: | EN |
| Publicado: |
Frontiers Media S.A.
2021
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/945aae3dad994147b8c34a2fa781bd23 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| id |
oai:doaj.org-article:945aae3dad994147b8c34a2fa781bd23 |
|---|---|
| record_format |
dspace |
| spelling |
oai:doaj.org-article:945aae3dad994147b8c34a2fa781bd232021-11-30T14:13:08ZDeep Learning-Augmented Head and Neck Organs at Risk Segmentation From CT Volumes2296-424X10.3389/fphy.2021.743190https://doaj.org/article/945aae3dad994147b8c34a2fa781bd232021-11-01T00:00:00Zhttps://www.frontiersin.org/articles/10.3389/fphy.2021.743190/fullhttps://doaj.org/toc/2296-424XPurpose: A novel deep learning model, Siamese Ensemble Boundary Network (SEB-Net) was developed to improve the accuracy of automatic organs-at-risk (OARs) segmentation in CT images for head and neck (HaN) as well as small organs, which was verified for use in radiation oncology practice and is therefore proposed.Methods: SEB-Net was designed to transfer CT slices into probability maps for the HaN OARs segmentation purpose. Dual key contributions were made to the network design to improve the accuracy and reliability of automatic segmentation toward the specific organs (e.g., relatively tiny or irregularly shaped) without sacrificing the field of view. The first implements an ensemble of learning strategies with shared weights that aggregates the pixel-probability transfer at three orthogonal CT planes to ameliorate 3D information integrity; the second exploits the boundary loss that takes the form of a distance metric on the space of contours to mitigate the challenges of conventional region-based regularization, when applied to highly unbalanced segmentation scenarios. By combining the two techniques, enhanced segmentation could be expected by comprehensively maximizing inter- and intra-CT slice information. In total, 188 patients with HaN cancer were included in the study, of which 133 patients were randomly selected for training and 55 for validation. An additional 50 untreated cases were used for clinical evaluation.Results: With the proposed method, the average volumetric Dice similarity coefficient (DSC) of HaN OARs (and small organs) was 0.871 (0.900), which was significantly higher than the results from Ua-Net, Anatomy-Net, and SRM by 4.94% (26.05%), 7.80% (24.65%), and 12.97% (40.19%), respectively. By contrast, the average 95% Hausdorff distance (95% HD) of HaN OARs (and small organs) was 2.87 mm (0.81 mm), which improves the other three methods by 50.94% (75.45%), 88.41% (79.07%), and 5.59% (67.98%), respectively. After delineation by SEB-Net, 81.92% of all organs in 50 HaN cancer untreated cases did not require modification for clinical evaluation.Conclusions: In comparison to several cutting-edge methods, including Ua-Net, Anatomy-Net, and SRM, the proposed method is capable of substantially improving segmentation accuracy for HaN and small organs from CT imaging in terms of efficiency, feasibility, and applicability.Wei WangWei WangQingxin WangQingxin WangQingxin WangMengyu JiaZhongqiu WangChengwen YangChengwen YangDaguang ZhangDaguang ZhangShujing WenDelong HouNingbo LiuNingbo LiuPing WangPing WangJun WangJun WangFrontiers Media S.A.articleradiotherapyconvolutional neural networksautomatic segmentationhead and neck cancerdeep learningPhysicsQC1-999ENFrontiers in Physics, Vol 9 (2021) |
| institution |
DOAJ |
| collection |
DOAJ |
| language |
EN |
| topic |
radiotherapy convolutional neural networks automatic segmentation head and neck cancer deep learning Physics QC1-999 |
| spellingShingle |
radiotherapy convolutional neural networks automatic segmentation head and neck cancer deep learning Physics QC1-999 Wei Wang Wei Wang Qingxin Wang Qingxin Wang Qingxin Wang Mengyu Jia Zhongqiu Wang Chengwen Yang Chengwen Yang Daguang Zhang Daguang Zhang Shujing Wen Delong Hou Ningbo Liu Ningbo Liu Ping Wang Ping Wang Jun Wang Jun Wang Deep Learning-Augmented Head and Neck Organs at Risk Segmentation From CT Volumes |
| description |
Purpose: A novel deep learning model, Siamese Ensemble Boundary Network (SEB-Net) was developed to improve the accuracy of automatic organs-at-risk (OARs) segmentation in CT images for head and neck (HaN) as well as small organs, which was verified for use in radiation oncology practice and is therefore proposed.Methods: SEB-Net was designed to transfer CT slices into probability maps for the HaN OARs segmentation purpose. Dual key contributions were made to the network design to improve the accuracy and reliability of automatic segmentation toward the specific organs (e.g., relatively tiny or irregularly shaped) without sacrificing the field of view. The first implements an ensemble of learning strategies with shared weights that aggregates the pixel-probability transfer at three orthogonal CT planes to ameliorate 3D information integrity; the second exploits the boundary loss that takes the form of a distance metric on the space of contours to mitigate the challenges of conventional region-based regularization, when applied to highly unbalanced segmentation scenarios. By combining the two techniques, enhanced segmentation could be expected by comprehensively maximizing inter- and intra-CT slice information. In total, 188 patients with HaN cancer were included in the study, of which 133 patients were randomly selected for training and 55 for validation. An additional 50 untreated cases were used for clinical evaluation.Results: With the proposed method, the average volumetric Dice similarity coefficient (DSC) of HaN OARs (and small organs) was 0.871 (0.900), which was significantly higher than the results from Ua-Net, Anatomy-Net, and SRM by 4.94% (26.05%), 7.80% (24.65%), and 12.97% (40.19%), respectively. By contrast, the average 95% Hausdorff distance (95% HD) of HaN OARs (and small organs) was 2.87 mm (0.81 mm), which improves the other three methods by 50.94% (75.45%), 88.41% (79.07%), and 5.59% (67.98%), respectively. After delineation by SEB-Net, 81.92% of all organs in 50 HaN cancer untreated cases did not require modification for clinical evaluation.Conclusions: In comparison to several cutting-edge methods, including Ua-Net, Anatomy-Net, and SRM, the proposed method is capable of substantially improving segmentation accuracy for HaN and small organs from CT imaging in terms of efficiency, feasibility, and applicability. |
| format |
article |
| author |
Wei Wang Wei Wang Qingxin Wang Qingxin Wang Qingxin Wang Mengyu Jia Zhongqiu Wang Chengwen Yang Chengwen Yang Daguang Zhang Daguang Zhang Shujing Wen Delong Hou Ningbo Liu Ningbo Liu Ping Wang Ping Wang Jun Wang Jun Wang |
| author_facet |
Wei Wang Wei Wang Qingxin Wang Qingxin Wang Qingxin Wang Mengyu Jia Zhongqiu Wang Chengwen Yang Chengwen Yang Daguang Zhang Daguang Zhang Shujing Wen Delong Hou Ningbo Liu Ningbo Liu Ping Wang Ping Wang Jun Wang Jun Wang |
| author_sort |
Wei Wang |
| title |
Deep Learning-Augmented Head and Neck Organs at Risk Segmentation From CT Volumes |
| title_short |
Deep Learning-Augmented Head and Neck Organs at Risk Segmentation From CT Volumes |
| title_full |
Deep Learning-Augmented Head and Neck Organs at Risk Segmentation From CT Volumes |
| title_fullStr |
Deep Learning-Augmented Head and Neck Organs at Risk Segmentation From CT Volumes |
| title_full_unstemmed |
Deep Learning-Augmented Head and Neck Organs at Risk Segmentation From CT Volumes |
| title_sort |
deep learning-augmented head and neck organs at risk segmentation from ct volumes |
| publisher |
Frontiers Media S.A. |
| publishDate |
2021 |
| url |
https://doaj.org/article/945aae3dad994147b8c34a2fa781bd23 |
| work_keys_str_mv |
AT weiwang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT weiwang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT qingxinwang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT qingxinwang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT qingxinwang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT mengyujia deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT zhongqiuwang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT chengwenyang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT chengwenyang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT daguangzhang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT daguangzhang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT shujingwen deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT delonghou deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT ningboliu deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT ningboliu deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT pingwang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT pingwang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT junwang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes AT junwang deeplearningaugmentedheadandneckorgansatrisksegmentationfromctvolumes |
| _version_ |
1718406518531948544 |